Over the past couple of years, Rational has modernized its client-side developer tools under the Eclipse framework. It came after some false starts, such as XDE, which was supposed to be the successor to the Rational Rose modeling tool, and Rational Rapid Developer.
Last year saw the release of testing tools that linked to Tivoli for closing the loop on performance testing with actual production data, the first upgrades to Rational Portfolio Manager (RPM), and the long-awaited open source updating of Rational’s process tools, which will also cover alternatives to Rational’s Unified Process (RUP).
And last year saw the changing of the guard. Danny Sabbah, whose background included chief technology architect for WebSphere, succeeded Rational co-founder Mike Devlin. To many observers, the transition couldn’t have happened soon enough because, until recently, many felt that Rational’s best days were behind it.
Way back when, Rational invented the concept of a unified software development lifecycle. It spanned use case modeling and requirements management to architecture, design, development, and testing. A decade ago, it hired three of the biggest names in the object-oriented development world, who created the language that became UML.
And in the couple of years following UML, Rational embarked on an acquisition frenzy that brought in tools covering all parts of development, design, and testing. Only today are players like Borland, Compuware, Serena, and Telelogic catching up in breadth of offerings.
The only problem was that Rational largely sat on its acquisitions. Taking its eye off the prize, Rational linked, but didn’t fully integrate, the tools or move to common engines. Instead, it embarked on wild forays aimed at cultivating new market segments.
Some of the failures were comical, such as Content Studio, which married the ClearCase source code control with Vignette’s content management tooling. The notion of creative staff acting like software developers was a culture chasm too difficult to cross.
Others were more damaging, such as the abortive Catapulse spinoff in 2000. There, several Rational founders tried inventing their own version of the CollabNet hosted development portal. When the venture fizzled barely a year or so later, Rational bought back the assets, sucking away many millions of dollars.
While Mr Sabbah’s background is clearly software engineering, his goal at Rational is to take the tooling framework beyond the software engineering organization. Although the technology was already well under development by the time Mr Sabbah assumed the reins, the cross-fertilization with Tivoli could be considered one of the first fruits.
Of course, with IBM Software always having promoted the cross-fertilization of technologies between brands, there have been questions as to whether Rational would survive or differentiate itself from WebSphere, now that a WebSphere veteran has taken the helm. For example, Rational’s IDE is essentially a rebranding of WebSphere Studio Application Developer (WSAD).
Mr Sabbah’s answer is that Rational products will continue focusing on development while WebSphere focuses on deployment and integration. Rational’s strategy involves supporting the kind of distributed development that is commonplace for larger software organizations today. More importantly, it involves closing the loop with the operations side and the business in order to provide definitive answers on how well software is supporting the enterprise.
Of course, Rational is not alone in that quest. Mercury, which acquired Kintana, is at least 12 months ahead of IBM in project portfolio management (PPM), the tooling and process intended as a mix of ERP and BPM dashboard for IT. Others such as Compuware, Serena, Computer Associates, and Telelogic have made similar acquisitions.
But the real riddle isn’t how well Rational will compete in the new PPM space, but whether it can revive the team tools that are its crown jewels. Today, ClearCase, ClearQuest, and RequisitePro are 10-year old products that continue to run on their own separate engines. Rational’s answer to what’s coming next in team tools will answer the question as to whether this IBM unit has clearly turned the corner.
Admittedly, Rational can be justly criticized for not having addressed this issue earlier. However, unifying tooling for the application life cycle has proven a riddle all along. Otherwise, Rational, Borland, Telelogic and others would have solved it by now.
The hang-up is that developers don’t need to see the same information, in the same shell, with the same degree of timeliness as testers, business analysts, or program managers, and vice versa. The idea of a single common engine for the application life cycle is as laughable today as it was when CASE (computer-aided software engineering) was proposed in the 1980s.
Rational’s latest releases of its process tooling, and the links to its project portfolio management dashboard, provide hints that next-generation tools will be loosely coupled, using services oriented architectures, web services standards, and the Eclipse framework. If you recall that one of Mr Sabbah’s previous roles at WebSphere was SOA evangelist, the hints of how the pieces will fit together in the Rational stable start to become clear.
